Open AccessApplication of GA-BP neural network model for small watershed flood forecasting in Chun’an county, China
Author(s) -
Dayong Huang,
Chongguo Tian,
Jiale Jiang
Publication year - 2020
Publication title -
iop conference series. earth and environmental science
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.179
H-Index - 26
eISSN - 1755-1307
pISSN - 1755-1315
DOI - 10.1088/1755-1315/612/1/012066
Subject(s) - artificial neural network , flood myth , watershed , flood forecasting , correlation coefficient , genetic algorithm , backpropagation , hydrology (agriculture) , environmental science , china , surface runoff , computer science , meteorology , data mining , geology , artificial intelligence , geography , machine learning , ecology , geotechnical engineering , archaeology , biology
Flood forecasting for small basins in hilly areas is often plagued by poor performance of hydrological models due to lack of observed data, meanwhile, the traditional Back Propagation (BP) neural network is easy to fall into the local minimum. This paper put forward an approach combined Genetic Algorithms (GA) with BP neural network and established a GA-BP neural network model to promote the flood forecasting. The flood hygrograph of Fenglingang small watershed, in Chun’an county, simulated by GA-BP model indicates that the deviation of runoff volumes is controlled within 10%, the deviation of peak discharge is kept below 20%, and absolute error of time to peak is less than 2h. Additionally, the correlation coefficient of simulation result of GA-BP model for each rainstorm event is above 0.75, which is smaller than that of traditional BP model. Consequently, it is demonstrated that the GA-BP model has a higher simulation precision and can provide reference for local forecasting in the future.